I hope that in Linus' upcoming debate against the representatives of DH Brown and Microsoft (isn't that happening in conjunction with Spring Comdex?), that he can set them straight on a few things.
From what I've heard from a friend of mine who as a dual CPU machine, Linux 2.0.x kernels was on par with NT 4.0 for scalability. With the 2.2.x kernels, Linux scales noticeably better. I don't know anyone who personally has a quad processor Linux box, but I've seen personally that NT 4.0 doesn't seem to get big improvements in performance going between dual and quad processors and the anecdotal evidence I've seen is that Linux 2.0.x also scales better on four processors than NT 4.0 and that Linux 2.2.x is an even bigger improvement over 2.0.x for four processors than for two.
DH Brown complains that there aren't a lot of published comparative benchmarks for Linux. But they are supposed to be a research organization. Why didn't they test it themselves? Why does news.com echo such uncredible criticism without question?
Instead they seem to increase the confusion by mixing the comparison between high-end *nixes that at this point are with little doubt more scalable than Linux (although Linux is gaining), with some of the feature comparisons (memory address size and journaled file system) of NT to imply that NT is also more scalable than Linux.
As many people have probably already noted, a journaling file system is already under development for Linux (I believe it is being written by Stephen Tweedie). From my experience with NT, I'd have to say that its 'journaling' file system is certainly not on par with AIX's JFS or Veritas on Solaris either. I've lost data on NTFS due to corruption. I've never lost data under ext2 except when I've had the whole hard drive fail. I've also seen NT take as long or longer doing 'file system checks' which theoretically shouldn't be necessary under a journaled file system. I certainly have never seen that happen with AIX's JFS or Veritas on Solaris.
As for the memory limitations, they just aren't a big deal for most applications. Linux is certainly on par with NT Server 4.0 here, and in all reality with a kernel recompile is on par with NT 4.0 Enterprise Edition. Frankly more than 2G of RAM (or 3G with a recompile) is about all that is realistically possible on x86 hardware. NT can't even take real advantage of 64-bit hardware yet, which is an area where Linux beats it hands down. Sure, the commercial *nixes beat Linux, but even NT 4.0 Enterprise Edition is an order of magnitude more expensive than even the most expensive Linux distribution without even comparing client licenses.
I wish I had an SMP machine so I could run and/or write the benchmarks myself.
Would I like to see what few limitations are currently in Linux lifted? Sure. Do they really matter to most people? No. Do I believe that the limitations in Linux will be lifted before the limitations in NT are fixed? Yes.
The tone of the news.com article is unecessarily negative when it doesn't need to be. An article which concludes that Linux nearly matches costly solutions for a fraction of the price could easily have been written from the same (few) data points that were presented in the article.
No kidding... This article echos claims that Linux doesn't support SMP or syslog.
It also says that Linux is less suitable for general business purposes than any of the commercial UNIXes or even NT, without giving much of any details on what criteria were used to make that determination.
However, you can go get numerous studies telling you that NT is cheaper to support than either Novell or Unix, by a significant amount.
I haven't seen one yet that I couldn't easily find outright bias and/or gross inaccuracies in. Some of them looked like they were directly paid for by Microsoft.
Can anyone point to a study that indicates supporting Linux is cheaper than supporting NT? Maybe for a web server or a DNS server, but how about for a complex environment (file and print, mail, database) across multiple sites? Again, I'm not sure - I'd like to see some data?
You might read the Kirch paper and the links from it. Because the big consulting companies have only recently noticed Linux's existance, they haven't done much serious research. What I have seen them do is often biased or otherwise flawed due to their lack of knowledge and/or understanding of Linux and *nix in general. My experience is that Linux in general has similar support issues to most commercial *nixes, and that they generally require considerably less support per box and less boxes for the same work than NT. For file and print I know from experience that *nix with Samba works better and is more reliable than NT. For email, its no comparison, especially for large installations. I know people who support 3000 POP/IMAP mail clients on a single processor Sparc 10 Model 50 with 128M of RAM and a few quad 18G drive packs attached. NT with Exchange server struggles to handle 300 to 500 Outlook users on a Quad Xeon box with 256M of RAM (I've seen it bog down to the point where delivery of email is delayed by hours). Furthermore Exhange server isn't at all reliable. It crashes and loses data occasionally, and requires almost weekly reboots. About the only good thing you can say about it is that it is a significant improvement over its predecessor MS-Mail.
I'd like to see some reasonable and unbiased studies done, but given the monetary motivations and conflicts of interest with most of the big consulting companies, I am not holding my breath.
Do you really think a Unix administrator is the same price as an NT administrator with similar skills? Call your local temp agency and find out. I have -- Unix admins are 2x the rate of NT admins.
I don't know about temp agencies being a good source. I know for permanent employment, similarly skilled and experienced *nix and NT administrators are similarly priced, at least around here. I'd be afraid that with temp agencies you might not be getting what you are paying for with NT administrators. There are a lot of pretenders and wannabes in the NT world. It seems like any kid with 2 weeks of training class under their belt is being flogged off by headhunters as an NT administrator. It is much harder to fake *nix experience.
Right. Companies don't like paying for "service and support".
You sure could fool me on that one. I've seen companies pay through the nose for support for software they didn't even need support for. Companies are paranoid to the point they will pay whatever cost just to try to cover their @$$es. They will pay through the nose for support that is nearly worthless, just to have an 800 number they can shift any blame to.
For every $500 NT licence that a company buys, it spends thousands of dollars supporting that NT box.
Then there are all the client licenses they need to buy, and all the add-on software they need. The NT license is cheaper than the add-on software. The client licenses quickly add up to more than the initial NT license.
Linux advocates may scoff at the "supportablity" of NT, but one of the reasons that NT has taken over the workgroup server market its that it is supposedly much cheaper to support than Novell or Unix.
Supposedly is the key word here. Its a huge marketing lie. NT is more expensive to support than Novell or *NIX. NT has taken over its share of the workgroup server market because of marketing, marketing, marketing.
Linux has solved the $500 software licence problem, but I'm not sure they've solved the $50,000 "service and support" issue.
From my experience NT is more expensive to support because it just plain breaks down more often. Not to mention that you seem to need three to five times the number of NT servers to replace Novell for file/print services or *NIX for any purpose. It also takes more administrators for a given number of NT servers than Novell or *NIX because contrary to popular belief forcing a GUI for every sysadmin change slows the process down, and for the most part remote administration of NT is still not quite there. Microsoft's marketing tells people that they don't need professional administrators for NT, which isn't true. And when people figure out that they do need full time administrators, then they are told they can hire any bozo off the street and send him to a little training and pay them $25,000 a year. The truth is that you can't get by with that, and skilled NT administrators (the ones who can make it run even close to reliably) are just as expensive as *NIX administrators.
People want "easier" and "cheaper", in most cases much more than they want "better".
Too many people are penny wise and pound foolish. They don't look at the big picture when determining what is "cheaper". Too often people confuse "cheaper" with "less expensive" - there is a big difference. Even people who should know better are often fooled by things that seem easier at the cosmetic level while they are more complex (in the case of NT, unnecessarily so) underneath.
This only applies to the Full copy. Win98's upgrade copy does not include this floppy. It also (irritatingly) asks for proof of some sort of prior Windows product. Thank god I kept my WfWG install floppies. Apparently, that boot floppy costs about $100, given the difference in price of the two editions.
And try to find the "full copy". Very few stores around here seem to carry anything but the upgrade edition. Its also getting virtually impossible to buy Windows 95 at all anymore, as all most stores have these days seem to be 98.
Tripple Ha! I've run a large amount of my Apple II software under Linux using a free emulator, much of which predates the existance of the IBM PC, let alone Windows by a couple of years.
I tried something similar with a friend of mine who is a mainframer with some MS-DOS and MS-Windows experience but almost no Linux or UNIX experience at all. You'd think this would give Windows 95 an unfair advantage, but read on.
Given a properly configured new machine with a blank hard drive and nothing but a boxed Windows 95 and a boxed Official Red Hat Linux he was eventually able to get both installed successfully. However, the route he ended up to get there is not what one might expect.
He first attempted to install Windows 95, but failed because he couldn't figure out how to get any MS-DOS boot floppy he had to recognize the (generic Toshiba ATAPI) CD-ROM drive (his previous machine has a proprietary Mitsumi CD-ROM in it) and he didn't have the correct driver diskette for ATAPI or instructions on how to configure MSCDEX.EXE.
Red Hat, on the other hand, provided a boot floppy which recognized his CD-ROM drive and let him partition the hard drive. Once he had Linux installed, he used it to copy the Windows 95 CD onto the FAT partition he had created on the hard drive using Linux and was able to boot from an MS-DOS boot floppy and run the Windows 95 install from the hard drive.
My conclusion is that even for reasonably competent people, Windows 95 installation isn't significantly easier than Linux. Given properly configured hardware (as apparently the Slate authors did not have), the Red Hat boot floppies will normally detect CD-ROM drives and install with very little trouble.
FUD is a subset of propaganda. There are lots of types of propaganda that are not FUD. Microsoft, like most other companies uses the other types of propaganda in their advertising and PR. Not every company uses FUD, and of those that do some rely on it more and/or are better at it than others. IBM was the past master of FUD (the term was repudately coined one of IBM's mainframe competitors in the 70's). Microsoft relies heavily on FUD, but their record in using it successfully seems to be weaker than IBMs was.
The article was talking about NT Workstation which is not legal to use as a file server for more than a few clients. If you want to do file serving, you are supposed to buy NT Server, which is two to three times the price of NT Workstation, and an order of magnitude more expensive than The Official Red Hat Linux box (which is more expensive than the CheapBytes Red Hat CD by an order of magnitude). And that is before you pay for client licenses, each of which for NT server are close to the price of the Official Red Hat boxed edition.
NT does ship with Internet Information Server, which lets you run HTTP and FTP services out of the box
Again, this only applies to NT Server. Microsoft contends that it is not even legal to use a 3rd party web server package such as O'Reilly's with NT Workstation.
Where's the benchmark?
on
Few Quickies
·
· Score: 1
I suspect that's the real point of the bet. MS publishing a side by side comparison with Oracle would be worth a million dollars to Oracle.
Undoubtedly, what I can't believe is that Microsoft was foolish enough to take the bait. For a company that has the track record they do of making smart marketing decisions, this seems like a serious error in judgement on their part.
Where's the benchmark?
on
Few Quickies
·
· Score: 1
So where's the benchmark of Microsoft coming close to Oracle's performance? Weren't they supposed to publish it to get the reward?
One would suspect that if Microsoft is reluctant to publish the results that while they might be more than 1/100th as fast, they might not be more than 1/16th as fast, meaning in all reality, it is still a 'win' for Oracle, as Microsoft has stated their solution is 1/16th the cost of the Oracle one. If they don't beat 1/16th the performance (which I believe unlikely given the differences in scaling efficiency between Solaris and NT, and the fact that the Sun Ultra Enterprise 10000 that is used in Oracle's test has 8 times as many processors as any box currently NT runs on, as well as the advantage of being 64 bit and having a considerably better I/O design than anything NT currently runs on.
I am not an expert on the subject, but I would say that reliability matters a lot in this situation.
I would concur and further suggest that one doesn't need to be a technical expert to make that determination. Simple business common sense would dictate that decision.
The article mentions that NT + MS SQL were pitted against AIX + DB2.
Actually I believe it was an AS/400 which runs OS/400. It was also a single processor AS/400 compared to a multiprocessor NT box.
While NT outperformed IBM's solution in most tests, it crashed under heavy load. Now what were you saying about the cost per transaction?
Cost per transaction matters little if you can't reliably do a given load of transactions.
Microsoft's long term road map is related to seeing how long they can milk their cash cows by slowly doleing out minor enhancements and bug fixes. They will let others innovate, then they will immitate or purchase then proclaim the immitation to be the original or their purchase to be their idea all along.
The fact that Microsoft has a long range plan is one of their big lies. The original edition of "The Road Ahead" barely mentioned the Internet because they had no long term plan for it, they were betting on their failed "blackbird" AOL-like technology. Their revisionism led to them tacking on a chapter for the second edition.
Furthermore this story is a bit off-base, as his stories often are.
Biased Nick Petreley may be, off-base, I would dispute. For the most part, he tends to present his opinions in a pretty reasonably justified manner.
Is Linux considered "splintered" because it comes in 64-bit flavours? Of course not, so why should Windows be?
Because with *NIX the API stays the same regardless of the word size. *NIX didn't have a huge crisis when the world moved from 16 to 32 bit like Microsoft has gone through moving from Win16 to Win32. Microsoft has always writen their APIs too close to the processor, which has hurt them in the long run.
And surely the ability of some future Win64 to emulate 32 bit code should not be counted as detrimental.
The fact that it has to rely on emulation when it will be running on a platform that is supposed to provide hardware level instruction set compatibility is what is detrimental. The fact that NT on Alpha has failed to take advantage of the Alpha's 64 bit architecture for several years is detrimental.
Also, counting the numbers of machines running various Web servers and operating systems and attached to the Internet can't be extrapolated to, say, what IT people like to use generally. So let's not get carried away.
So what numbers do you suggest Petreley should have used instead? What is being used on publically accessable sites is certainly significant. Its hard to say whether any disparity between what IT people like to use in general and what they are deploying on the Internet would fall in favor or against Microsoft.
Anyway, what I'd really like to know is how many big NT sites use it because they receive "incentives" from Microsoft in cash and kind. At least, four years ago this was happening at some companies I was close to
I would suspect that it is still going on. Even if that sort of coercion isn't, it is hard to dispute that the decision to deploy NT in many IT shops is made by nontechnical managers for political reasons rather than by technical people for technical or financial reasons.
Be honest now, if Hotmail used Linux instead of FreeBSD for their web servers, do you think Solaris would have been mentioned at all?
Yes, because the typical PHB target audience is familiar with the Sun-Microsoft lawsuits and general animosity recently, so the fact that Microsoft depends on Sun's products for one of their services is a significant fact.
What do birds commonly found near large bodies of water have to do with anything? Seriously, I think you mean "gall".:-)
Seriously though, why do you think NT needs third party "evangelists"? Microsoft has oodles to spend on advertising. Evangelism is for those that don't have huge advertising and PR budgets.
Or maybe they might also be people with a monetary reason to want to see Amazon sell copies of the book (like they own Amazon stock or something). If its not one ulterior motive, its another.:-)
Going back to the topic of this whole posting, maybe that's a good machine to try out Slackware 4.0 on.:)
Perhaps, although I am happy with SuSE on it at the moment, albiet I probably should upgrade it to 6.0 or 6.1. I don't have any experience with that model, but Debian is probably easier to get installed, it seems to give a lot more flexibility in the install process.
I don't have any recent versions of Debian, nor do I have a fast enough net connection to find downloading it pleasant. Someday I should at least try out Debian, it is the only major distro I haven't ever played with.
I assume the system doesn't have a CD-Rom,
You are correct. The first install I did on the machine was Slackware 3.4 which I did from floppies I made from a CD on another machine. That is very time consuming, but not difficult. Later I purchased a new hard drive (2.1G to replace the 270M drive it came with), a 32M memory card (to replace the 8M card that came in it and expand it from 12M to 36M), and a MegaHertz PCMCIA combo 33.6 modem and Ethernet. I attempted to install Red Hat, but as I mentioned, due to the floppy drive funkiness on this series of ThinkPads (which apparent afflicts the 75x series as well as the 35x series), that failed. However, as I mentioned, I was able to get a network install of SuSE 5.3 to work, so all is happy.
but if you can boot to DOS on it, you could install the packages you need on a DOS/umsdos partition and try booting it from loadlin.
I don't use MS-DOS (for that matter I don't even think I have a copy of it anywhere), so that isn't really an option. I was told the machine had MS-DOS 6.x and Windows 3.1x on it when I got it, but I never booted it that way to find out, the first thing I did was boot Slackware and fdisk away the Windows infestation. I do have a copy of DR-DOS 6.0 around somewhere, so I suppose I could have played with that. However, due to limited disk when I first purchased the machine and no network connection, that didn't seem very workable.
But the idea that RedHat can't install onto laptops is just silly.
Very true. Unfortunately, as I noted, getting any OS installed on some laptops can be an adventure. For me it was easier to just use SuSE, since I already had it, and it worked without hassles. It is not completely fair to pick on Red Hat because of the problems some people will encounter with laptops (and that isn't what my intention was), as it is mostly the hardware manufacturer's fault. At any rate, I personally would guess getting Windows 95/98 installed on this laptop would be more difficult than SuSE or Slackware. Thankfully, since I don't have (or want) Windows 95/98, that isn't an issue.
Too bad you'll never see the SuSE box here in the States
They've got it at the local Borders, and I live in a small (~300,000) midwestern town. The same Borders also carries Red Hat, Red Hat Power Tools, SuSE Linux Office 99, Caldera OpenLinux and quite a number of Linux books. Viva La Borders!
I'm curious about those versions of redhat that don't install on laptops
Well, I have an IBM ThinkPad 355Cs (which is an old and slightly funky semi-proprietary 486SX-33 laptop). I can't get it to read the Red Hat 5.1 or 5.2 boot floppy. I've been able to get both Slackware 3.4 and SuSE 5.3 installed on it. I've tried making Red Hat boot floppies on several machines (thinking it is an alignment problem or something). This model ThinkPad requires "floppy=thinkpad" as boot line parameter or an append with Slackware and SuSE. The Red Hat boot floppy doesn't seem to grok that or something.
Laptops aren't normal x86 boxes though... they are often highly funky. It does sometimes take a considerable amount of experimentation to get them to play nicely, even with Windows.
I know Linux on a 4 meg system may not be reasonable any more,
I've installed Slackware 3.4 on a 386SX-25 with 4M from a local CD-ROM drive (1X no less). I haven't tried any more recent versions yet, as I rarely use that machine (for obvious reasons:-) ).
but it seems 8 should be doable, but 8's too small to install RedHat,
That is not necessarily true. I have successfully installed Red Hat 5.2 (via NFS and Ethernet) on a 486DLC-40 (similar performance wise to a 486DX-33) with 8M. It is slow, especially when running X, but it does work reliably. One of these days I will have to get around to bumping the memory in that box up to 20M or 32M.
First, why do you need my "REAL NAME"? Are you going to sue me? Why do you even care? Why can't I have multiple identities?
It is virtually impossible to make sure you have "real names", even for a subscription service. It is too easy to falsify an identity, particularly online.
Second, charge me $29 and I'll go away. I'll bet most everybody will go away. Has been tried, doesn't work.
Pay services on that level are a hard sell unless they have really compelling content that is exclusive. A hard thing to come by. There is just too much content available on the web for free for sites to get away with charging that kind of subscription fee. Some people even whined when Slashdot added banner ads.
Third, DejaNews makes Usenet quite usable.
Well, I used to be an avid USENET reader/poster. Over the past few years I don't read it like I used to. Partially it is because I don't have enough network bandwidth to make it pleasant, but it is also because the signal to noise ratio on USENET definitely went downhill. DejaNews is a really great service that makes finding the good content on USENET a lot easier, albiet it isn't the most convenient way just to browse news groups.
Slashdot Mirror
I hope that in Linus' upcoming debate against the representatives of DH Brown and Microsoft (isn't that happening in conjunction with Spring Comdex?), that he can set them straight on a few things.
From what I've heard from a friend of mine who as a dual CPU machine, Linux 2.0.x kernels was on par with NT 4.0 for scalability. With the 2.2.x kernels, Linux scales noticeably better. I don't know anyone who personally has a quad processor Linux box, but I've seen personally that NT 4.0 doesn't seem to get big improvements in performance going between dual and quad processors and the anecdotal evidence I've seen is that Linux 2.0.x also scales better on four processors than NT 4.0 and that Linux 2.2.x is an even bigger improvement over 2.0.x for four processors than for two.
DH Brown complains that there aren't a lot of published comparative benchmarks for Linux. But they are supposed to be a research organization. Why didn't they test it themselves? Why does news.com echo such uncredible criticism without question?
Instead they seem to increase the confusion by mixing the comparison between high-end *nixes that at this point are with little doubt more scalable than Linux (although Linux is gaining), with some of the feature comparisons (memory address size and journaled file system) of NT to imply that NT is also more scalable than Linux.
As many people have probably already noted, a journaling file system is already under development for Linux (I believe it is being written by Stephen Tweedie). From my experience with NT, I'd have to say that its 'journaling' file system is certainly not on par with AIX's JFS or Veritas on Solaris either. I've lost data on NTFS due to corruption. I've never lost data under ext2 except when I've had the whole hard drive fail. I've also seen NT take as long or longer doing 'file system checks' which theoretically shouldn't be necessary under a journaled file system. I certainly have never seen that happen with AIX's JFS or Veritas on Solaris.
As for the memory limitations, they just aren't a big deal for most applications. Linux is certainly on par with NT Server 4.0 here, and in all reality with a kernel recompile is on par with NT 4.0 Enterprise Edition. Frankly more than 2G of RAM (or 3G with a recompile) is about all that is realistically possible on x86 hardware. NT can't even take real advantage of 64-bit hardware yet, which is an area where Linux beats it hands down. Sure, the commercial *nixes beat Linux, but even NT 4.0 Enterprise Edition is an order of magnitude more expensive than even the most expensive Linux distribution without even comparing client licenses.
I wish I had an SMP machine so I could run and/or write the benchmarks myself.
Would I like to see what few limitations are currently in Linux lifted? Sure. Do they really matter to most people? No. Do I believe that the limitations in Linux will be lifted before the limitations in NT are fixed? Yes.
The tone of the news.com article is unecessarily negative when it doesn't need to be. An article which concludes that Linux nearly matches costly solutions for a fraction of the price could easily have been written from the same (few) data points that were presented in the article.
No kidding... This article echos claims that Linux doesn't support SMP or syslog.
It also says that Linux is less suitable for general business purposes than any of the commercial UNIXes or even NT, without giving much of any details on what criteria were used to make that determination.
NE1000
NE1000! Yikes. You poor SOD. An NE2000 is bad enough... The NE1000 is an 8 bit card if memory serves.
However, you can go get numerous studies telling you that NT is cheaper to support than either Novell or Unix, by a significant amount.
I haven't seen one yet that I couldn't easily find outright bias and/or gross inaccuracies in. Some of them looked like they were directly paid for by Microsoft.
Can anyone point to a study that indicates supporting Linux is cheaper than supporting NT? Maybe for a web server or a DNS server, but how about for a complex environment (file and print, mail, database) across multiple sites? Again, I'm not sure - I'd like to see some data?
You might read the Kirch paper and the links from it. Because the big consulting companies have only recently noticed Linux's existance, they haven't done much serious research. What I have seen them do is often biased or otherwise flawed due to their lack of knowledge and/or understanding of Linux and *nix in general. My experience is that Linux in general has similar support issues to most commercial *nixes, and that they generally require considerably less support per box and less boxes for the same work than NT. For file and print I know from experience that *nix with Samba works better and is more reliable than NT. For email, its no comparison, especially for large installations. I know people who support 3000 POP/IMAP mail clients on a single processor Sparc 10 Model 50 with 128M of RAM and a few quad 18G drive packs attached. NT with Exchange server struggles to handle 300 to 500 Outlook users on a Quad Xeon box with 256M of RAM (I've seen it bog down to the point where delivery of email is delayed by hours). Furthermore Exhange server isn't at all reliable. It crashes and loses data occasionally, and requires almost weekly reboots. About the only good thing you can say about it is that it is a significant improvement over its predecessor MS-Mail.
I'd like to see some reasonable and unbiased studies done, but given the monetary motivations and conflicts of interest with most of the big consulting companies, I am not holding my breath.
Do you really think a Unix administrator is the same price as an NT administrator with similar skills? Call your local temp agency and find out. I have -- Unix admins are 2x the rate of NT admins.
I don't know about temp agencies being a good source. I know for permanent employment, similarly skilled and experienced *nix and NT administrators are similarly priced, at least around here. I'd be afraid that with temp agencies you might not be getting what you are paying for with NT administrators. There are a lot of pretenders and wannabes in the NT world. It seems like any kid with 2 weeks of training class under their belt is being flogged off by headhunters as an NT administrator. It is much harder to fake *nix experience.
Right. Companies don't like paying for "service and support".
You sure could fool me on that one. I've seen companies pay through the nose for support for software they didn't even need support for. Companies are paranoid to the point they will pay whatever cost just to try to cover their @$$es. They will pay through the nose for support that is nearly worthless, just to have an 800 number they can shift any blame to.
For every $500 NT licence that a company buys, it spends thousands of dollars supporting that NT box.
Then there are all the client licenses they need to buy, and all the add-on software they need. The NT license is cheaper than the add-on software. The client licenses quickly add up to more than the initial NT license.
Linux advocates may scoff at the "supportablity" of NT, but one of the reasons that NT has taken over the workgroup server market its that it is supposedly much cheaper to support than Novell or Unix.
Supposedly is the key word here. Its a huge marketing lie. NT is more expensive to support than Novell or *NIX. NT has taken over its share of the workgroup server market because of marketing, marketing, marketing.
Linux has solved the $500 software licence problem, but I'm not sure they've solved the $50,000 "service and support" issue.
From my experience NT is more expensive to support because it just plain breaks down more often. Not to mention that you seem to need three to five times the number of NT servers to replace Novell for file/print services or *NIX for any purpose. It also takes more administrators for a given number of NT servers than Novell or *NIX because contrary to popular belief forcing a GUI for every sysadmin change slows the process down, and for the most part remote administration of NT is still not quite there. Microsoft's marketing tells people that they don't need professional administrators for NT, which isn't true. And when people figure out that they do need full time administrators, then they are told they can hire any bozo off the street and send him to a little training and pay them $25,000 a year. The truth is that you can't get by with that, and skilled NT administrators (the ones who can make it run even close to reliably) are just as expensive as *NIX administrators.
People want "easier" and "cheaper", in most cases much more than they want "better".
Too many people are penny wise and pound foolish. They don't look at the big picture when determining what is "cheaper". Too often people confuse "cheaper" with "less expensive" - there is a big difference. Even people who should know better are often fooled by things that seem easier at the cosmetic level while they are more complex (in the case of NT, unnecessarily so) underneath.
This only applies to the Full copy. Win98's upgrade copy does not include this floppy. It also (irritatingly) asks for proof of some sort of prior Windows product. Thank god I kept my WfWG install floppies.
Apparently, that boot floppy costs about $100, given the difference in price of the two editions.
And try to find the "full copy". Very few stores around here seem to carry anything but the upgrade edition. Its also getting virtually impossible to buy Windows 95 at all anymore, as all most stores have these days seem to be 98.
Double Ha!
Tripple Ha! I've run a large amount of my Apple II software under Linux using a free emulator, much of which predates the existance of the IBM PC, let alone Windows by a couple of years.
So what does FUD stand for?
Fear, Uncertainty and Doubt.
A particularly effective (if not properly countered) form of negative propaganda.
I tried something similar with a friend of mine who is a mainframer with some MS-DOS and MS-Windows experience but almost no Linux or UNIX experience at all. You'd think this would give Windows 95 an unfair advantage, but read on.
Given a properly configured new machine with a blank hard drive and nothing but a boxed Windows 95 and a boxed Official Red Hat Linux he was eventually able to get both installed successfully. However, the route he ended up to get there is not what one might expect.
He first attempted to install Windows 95, but failed because he couldn't figure out how to get any MS-DOS boot floppy he had to recognize the (generic Toshiba ATAPI) CD-ROM drive (his previous machine has a proprietary Mitsumi CD-ROM in it) and he didn't have the correct driver diskette for ATAPI or instructions on how to configure MSCDEX.EXE.
Red Hat, on the other hand, provided a boot floppy which recognized his CD-ROM drive and let him partition the hard drive. Once he had Linux installed, he used it to copy the Windows 95 CD onto the FAT partition he had created on the hard drive using Linux and was able to boot from an MS-DOS boot floppy and run the Windows 95 install from the hard drive.
My conclusion is that even for reasonably competent people, Windows 95 installation isn't significantly easier than Linux. Given properly configured hardware (as apparently the Slate authors did not have), the Red Hat boot floppies will normally detect CD-ROM drives and install with very little trouble.
Remember when it was called ''propaganda''?
FUD is a subset of propaganda. There are lots of types of propaganda that are not FUD. Microsoft, like most other companies uses the other types of propaganda in their advertising and PR. Not every company uses FUD, and of those that do some rely on it more and/or are better at it than others. IBM was the past master of FUD (the term was repudately coined one of IBM's mainframe competitors in the 70's). Microsoft relies heavily on FUD, but their record in using it successfully seems to be weaker than IBMs was.
But Jerk, what about file sharing? (duck)
The article was talking about NT Workstation which is not legal to use as a file server for more than a few clients. If you want to do file serving, you are supposed to buy NT Server, which is two to three times the price of NT Workstation, and an order of magnitude more expensive than The Official Red Hat Linux box (which is more expensive than the CheapBytes Red Hat CD by an order of magnitude). And that is before you pay for client licenses, each of which for NT server are close to the price of the Official Red Hat boxed edition.
NT does ship with Internet Information Server, which lets you run HTTP and FTP services out of the box
Again, this only applies to NT Server. Microsoft contends that it is not even legal to use a 3rd party web server package such as O'Reilly's with NT Workstation.
I suspect that's the real point of the bet. MS publishing a side by side comparison with Oracle would be worth a million dollars to Oracle.
Undoubtedly, what I can't believe is that Microsoft was foolish enough to take the bait. For a company that has the track record they do of making smart marketing decisions, this seems like a serious error in judgement on their part.
So where's the benchmark of Microsoft coming close to Oracle's performance? Weren't they supposed to publish it to get the reward?
One would suspect that if Microsoft is reluctant to publish the results that while they might be more than 1/100th as fast, they might not be more than 1/16th as fast, meaning in all reality, it is still a 'win' for Oracle, as Microsoft has stated their solution is 1/16th the cost of the Oracle one. If they don't beat 1/16th the performance (which I believe unlikely given the differences in scaling efficiency between Solaris and NT, and the fact that the Sun Ultra Enterprise 10000 that is used in Oracle's test has 8 times as many processors as any box currently NT runs on, as well as the advantage of being 64 bit and having a considerably better I/O design than anything NT currently runs on.
I am not an expert on the subject, but I would say that reliability matters a lot in this situation.
I would concur and further suggest that one doesn't need to be a technical expert to make that determination. Simple business common sense would dictate that decision.
The article mentions that NT + MS SQL were pitted against AIX + DB2.
Actually I believe it was an AS/400 which runs OS/400. It was also a single processor AS/400 compared to a multiprocessor NT box.
While NT outperformed IBM's solution in most tests, it crashed under heavy load. Now what were you saying about the cost per transaction?
Cost per transaction matters little if you can't reliably do a given load of transactions.
Please elaborate on the MS long term road map.
Microsoft's long term road map is related to seeing how long they can milk their cash cows by slowly doleing out minor enhancements and bug fixes. They will let others innovate, then they will immitate or purchase then proclaim the immitation to be the original or their purchase to be their idea all along.
The fact that Microsoft has a long range plan is one of their big lies. The original edition of "The Road Ahead" barely mentioned the Internet because they had no long term plan for it, they were betting on their failed "blackbird" AOL-like technology. Their revisionism led to them tacking on a chapter for the second edition.
Furthermore this story is a bit off-base, as his stories often are.
Biased Nick Petreley may be, off-base, I would dispute. For the most part, he tends to present his opinions in a pretty reasonably justified manner.
Is Linux considered "splintered" because it comes in 64-bit flavours? Of course not, so why should Windows be?
Because with *NIX the API stays the same regardless of the word size. *NIX didn't have a huge crisis when the world moved from 16 to 32 bit like Microsoft has gone through moving from Win16 to Win32. Microsoft has always writen their APIs too close to the processor, which has hurt them in the long run.
And surely the ability of some future Win64 to emulate 32 bit code should not be counted as detrimental.
The fact that it has to rely on emulation when it will be running on a platform that is supposed to provide hardware level instruction set compatibility is what is detrimental. The fact that NT on Alpha has failed to take advantage of the Alpha's 64 bit architecture for several years is detrimental.
Also, counting the numbers of machines running various Web servers and operating systems and attached to the Internet can't be extrapolated to, say, what IT people like to use generally. So let's not get carried away.
So what numbers do you suggest Petreley should have used instead? What is being used on publically accessable sites is certainly significant. Its hard to say whether any disparity between what IT people like to use in general and what they are deploying on the Internet would fall in favor or against Microsoft.
Anyway, what I'd really like to know is how many big NT sites use it because they receive "incentives" from Microsoft in cash and kind. At least, four years ago this was happening at some companies I was close to
I would suspect that it is still going on. Even if that sort of coercion isn't, it is hard to dispute that the decision to deploy NT in many IT shops is made by nontechnical managers for political reasons rather than by technical people for technical or financial reasons.
Be honest now, if Hotmail used Linux instead of FreeBSD for their web servers, do you think Solaris would have been mentioned at all?
Yes, because the typical PHB target audience is familiar with the Sun-Microsoft lawsuits and general animosity recently, so the fact that Microsoft depends on Sun's products for one of their services is a significant fact.
I can't believe the gull of some people.
:-)
What do birds commonly found near large bodies of water have to do with anything? Seriously, I think you mean "gall".
Seriously though, why do you think NT needs third party "evangelists"? Microsoft has oodles to spend on advertising. Evangelism is for those that don't have huge advertising and PR budgets.
must have been written by Micro$oft flaks
:-)
Or maybe they might also be people with a monetary reason to want to see Amazon sell copies of the book (like they own Amazon stock or something). If its not one ulterior motive, its another.
Going back to the topic of this whole posting, maybe that's a good machine to try out Slackware 4.0 on. :)
Perhaps, although I am happy with SuSE on it at the moment, albiet I probably should upgrade it to 6.0 or 6.1.
I don't have any experience with that model, but Debian is probably easier to get installed, it seems to give a lot more flexibility in the install process.
I don't have any recent versions of Debian, nor do I have a fast enough net connection to find downloading it pleasant. Someday I should at least try out Debian, it is the only major distro I haven't ever played with.
I assume the system doesn't have a CD-Rom,
You are correct. The first install I did on the machine was Slackware 3.4 which I did from floppies I made from a CD on another machine. That is very time consuming, but not difficult. Later I purchased a new hard drive (2.1G to replace the 270M drive it came with), a 32M memory card (to replace the 8M card that came in it and expand it from 12M to 36M), and a MegaHertz PCMCIA combo 33.6 modem and Ethernet. I attempted to install Red Hat, but as I mentioned, due to the floppy drive funkiness on this series of ThinkPads (which apparent afflicts the 75x series as well as the 35x series), that failed. However, as I mentioned, I was able to get a network install of SuSE 5.3 to work, so all is happy.
but if you can boot to DOS on it, you could install the packages you need on a DOS/umsdos partition and try booting it from loadlin.
I don't use MS-DOS (for that matter I don't even think I have a copy of it anywhere), so that isn't really an option. I was told the machine had MS-DOS 6.x and Windows 3.1x on it when I got it, but I never booted it that way to find out, the first thing I did was boot Slackware and fdisk away the Windows infestation. I do have a copy of DR-DOS 6.0 around somewhere, so I suppose I could have played with that. However, due to limited disk when I first purchased the machine and no network connection, that didn't seem very workable.
But the idea that RedHat can't install onto laptops is just silly.
Very true. Unfortunately, as I noted, getting any OS installed on some laptops can be an adventure. For me it was easier to just use SuSE, since I already had it, and it worked without hassles. It is not completely fair to pick on Red Hat because of the problems some people will encounter with laptops (and that isn't what my intention was), as it is mostly the hardware manufacturer's fault. At any rate, I personally would guess getting Windows 95/98 installed on this laptop would be more difficult than SuSE or Slackware. Thankfully, since I don't have (or want) Windows 95/98, that isn't an issue.
Too bad you'll never see the SuSE box here in the States
They've got it at the local Borders, and I live in a small (~300,000) midwestern town. The same Borders also carries Red Hat, Red Hat Power Tools, SuSE Linux Office 99, Caldera OpenLinux and quite a number of Linux books. Viva La Borders!
I'm curious about those versions of redhat that don't install on laptops
Well, I have an IBM ThinkPad 355Cs (which is an old and slightly funky semi-proprietary 486SX-33 laptop). I can't get it to read the Red Hat 5.1 or 5.2 boot floppy. I've been able to get both Slackware 3.4 and SuSE 5.3 installed on it. I've tried making Red Hat boot floppies on several machines (thinking it is an alignment problem or something). This model ThinkPad requires "floppy=thinkpad" as boot line parameter or an append with Slackware and SuSE. The Red Hat boot floppy doesn't seem to grok that or something.
Laptops aren't normal x86 boxes though... they are often highly funky. It does sometimes take a considerable amount of experimentation to get them to play nicely, even with Windows.
I know Linux on a 4 meg system may not be reasonable any more,
:-) ).
I've installed Slackware 3.4 on a 386SX-25 with 4M from a local CD-ROM drive (1X no less). I haven't tried any more recent versions yet, as I rarely use that machine (for obvious reasons
but it seems 8 should be doable, but 8's too small to install RedHat,
That is not necessarily true. I have successfully installed Red Hat 5.2 (via NFS and Ethernet) on a 486DLC-40 (similar performance wise to a 486DX-33) with 8M. It is slow, especially when running X, but it does work reliably. One of these days I will have to get around to bumping the memory in that box up to 20M or 32M.
First, why do you need my "REAL NAME"? Are you going to sue me? Why do you even care? Why can't I have multiple identities?
It is virtually impossible to make sure you have "real names", even for a subscription service. It is too easy to falsify an identity, particularly online.
Second, charge me $29 and I'll go away. I'll bet most everybody will go away. Has been tried, doesn't work.
Pay services on that level are a hard sell unless they have really compelling content that is exclusive. A hard thing to come by. There is just too much content available on the web for free for sites to get away with charging that kind of subscription fee. Some people even whined when Slashdot added banner ads.
Third, DejaNews makes Usenet quite usable.
Well, I used to be an avid USENET reader/poster. Over the past few years I don't read it like I used to. Partially it is because I don't have enough network bandwidth to make it pleasant, but it is also because the signal to noise ratio on USENET definitely went downhill. DejaNews is a really great service that makes finding the good content on USENET a lot easier, albiet it isn't the most convenient way just to browse news groups.
I couldn't have said it better myself. Some of these people are so two-faced.
I think it is more telling that about the only really reliably worthwhile part of ZDNet and PC Week is usually Spencer F. Kat. A cartoon. Sad really.